The Reaction of Triphenylphosphine with Some Aromatic Amine

12 oz., on partially crystalline material which on crystallization from extraction in the usual manner, yielded 11 g. of gummy ethanol furnished 0.24 ...
1 downloads 0 Views 247KB Size
TRIPHENYLPHOSPHINE WITH AROMATIC ah^^^ OXIDES

March 20, 19.59

Benzene eluted 5 g. of non-crystalline gum; ethanol eluted partially crystalline material which on crystallization from ethanol furnished 0.24 g. of colorless crystals, m.p. 221°, [a]D -90" (95% ethanol, c 0.50). Anal. Calcd. for ClrH180a: C, 68.68; H, 6.92. Cdlcd. for C16H2001: C, 68.16; H, 7.63. Found: C, 68.18; H, 6.74. TALLAHASSEE, FLA.

F1exuosin.-H. fEexuosum was collected near Tallahassee in June, 1956. The air-dried whole plant, w t . 12 oz., on extraction in the usual manner, yielded 11 g. of gummy material which could not be induced to crystallize. Chromatography over 40 g. of alumina (eluate benzene) gave 7 g. of viscous material; the more polar fractions were highly colored and discarded. The fractions eluted with benzene were rechromatographed over freshly-ignited alumina.

[CONTRIBUTION FROM

THE

1483

DEPARTMENT O F CHEMISTRY, TEMPLE UXIVERSITP]

The Reaction of Triphenylphosphine with Some Aromatic Amine Oxides1 BY EDGAR HOWARD, JR.,

AXD

WILLIAMF. OLSZEWSKI

RECEIVED AUGUST11, 1968

A new method is presented for the deoxygenation of pyridine-N-oxides by the use of triphenylphosphine. The yields of the corresponding pyridine and triphenylphosphine oxide are good. The reaction between 4-nitropyridine-N-oxide and triphenylphosphine gave triphenylphosphine oxide as the only product which could be isolated; a kinetic study showed this reaction to be first order with respect to each of the reactants.

Pyridine-N-oxide has received considerable attention recently as an intermediate in the synthesis of pyridine derivative^.^,^ Substitution reactions in the pyridine ring are much more difficult than in the benzene ring owing to electronic effects exerted by the hetero nitrogen atom. If, however, pyridine is converted to the N-oxide, the over-all distribution of electrons is modified. The effect is an increased concentration of electrons a t the 2and 4-positions, facilitating electrophilic substitution in these positions. Deoxygenation of pyridine-N-oxides may be accomplished by a variety of methods. A nucleophilic reagent (Nu) may attack the oxygen atom of pyridine-N-oxide as3

phine oxide. The same authors reported that, in (CH8)gN +0

0 cy+

O-?Nu

Electron donors such as phosphorus trichloride can supply an electron pair to the oxygen atom and cause deoxygenation. The most common laboratory method for accomplishing the deoxygenation consists in refluxing a chloroform solution of the N-oxide with phosphorus trichloride. Our search for a reagent of wider applicability has led to the use of triphenylphosphine as the deoxygenating agent. Although triphenylphosphine is substantially stable in air, it is attacked by a wide variety of oxygen-containing compounds with the formation of triphenylphosphine oxide. This property of triphenylphosphine, and of phosphines in general, has been utilized in the removal of oxygen atoms of several substrate molecules.4 I t has been reported4 that aliphatic amine oxides, such as trimethylamine oxide, react quantitatively with tertiary phosphines in boiling glacial acetic acid to form the corresponding amine and phos(1) From a thesis submitted by William F . Olszewski in partial fulfillment of the requirements for t h e M . A . degree, Temple University, June, 1968, (2) E. Ochiai, J. O r g . Chem., 18, 534 (1953). (3) A R. Ratritzky, Quart. Revs. (London), 10, 395 (1956). (4) 1,. Horner and H. Hoffmann, Angew. Chem., 68, 480 (1956).

+ PRI

----f

(CHJsN

+ OPRI

contrast, the N-oxides of pyridine and quinoline were found to be quite stable toward both triphenylphosphine and triethylphosphine. It was observed that the ease of removal of the oxygen from amine oxides often decreases with the dipole moment of the amine oxide, e.g., trimethylamine oxide 5.04 D., pyridine-N-oxide 4.24 D. In our investigation it was discovered that triphenylphosphine will cause deoxygenation of a variety of aromatic amine oxides a t high temperatures

i 0 I, RI 11, R1 111, R1 IV, R1

= R2 = H = CH3, R2

H = H, R2 CH3 = H, Rz = OCHa

Quinoline-N-oxide is also smoothly deoxygenated by triphenylphosphine. The optimum method for performing this reaction is to heat the reactants in the absence of a solvent. A i temperatures considerably above 200' the amine distils from the reaction mixture in good yield in most cases. Triphenylphosphine oxide was isolated from the residue by recrystallization from a water-methanol solvent pair. I n the reactions of 4-methoxypyridine-N-oxide and quinoline-N-oxide the products were isolated by column chromatography. The only solvent found suitable for the reaction was triethylene glycol; however, there is no apparent advantage derived from its use. Observations of the reaction of triphenylphosphine with several amine oxides are summarized in Table I. Considerable difficulty was encountered in the reaction of triphenylphosphine with 4-nitropyridine-N-oxide. Upon heating the reactants in the absence of a solvent a vigorous exothermic reaction

EDGAR HOWARD, JR.,

1484

.IND

TABLE I REACTION OF TRIPHENYLPHOSPHINE WITH AMISE OXIDES Amine-Noxide

Solvent

Pyridine

Reaction teOmmp., C.

Yield of product, yo Amine

(CliH6)3-

PO

..

Chloroforma Ethanol (sa%)* Chlorobenzene' Dibutyl carhitnl Sone Triethylene glycol

Reflux Reflux Reflux ?x0-2:3:3 2.jO-2GO

89 8 8n.8

72.G 05 2

4-Picaline

None Triethylene glycol

275-280 "0-280

92.8 AG 3

m.1

2-Picoline

L-one Triethylene glycol

275-280 230-240

51.0 50.0

49.0

4-hrethoxypyridine

NoneC Triethylcne glycold

270-2iR 270-275

G1.O

83.0

Quinoline

None' Triethylene glycold

270-280 280-285

89.2

!I4

4-Nitropyridine

Chloroforme Dimethylformamidef

Reflux

Koneg

51.1

50.2

:

Reflux 130

.;x. 7

.. KOevidence of a reaction until the solvent was removed Triethylene glycol

a

mxxn

170-176

by distillation and a free flame applied. 77.8% of triphenylphosphine was recovered. Products were isolated by column chromatography. Attempted isolation of products by column chromatography was too slow to be feasible. e 78.870 of triphenylphosphine was recovered. f Ultraviolet studies revealed the presence of triphenylphosphine oxide, but not 4-nitropyridine. 0 Violent reaction was observed; triphenylphosphine oxide was isolated by column chromatography.

occurred, accompanied by the evolution of brown nitrous fumes. The only product that could be isolated from the reaction mixture was triphenylphosphine oxide. Investigation of the reaction mixture according to the method of den Hertogs showed that neither of the known decomposition products of 4-nitropyridine, N-(4'-pyridyl)-4-pyridone nor 4-pyridone, was present. When the reaction was repeated in refluxing chloroform and dimethylformamide the conditions were found to be too mild for the reaction to proceed to any considerable extent. ,4 kinetic study of the reaction between triphenylphosphine and 4-nitropyridine-N-oxide in triethylene glycol was performed using a spectrophotometric method. Ultraviolet spectra of the four components of the system were determined in absolute methanol, and it was found that a t 328 mp 4-nitropyridine-N-oxide had a strong, unique absorption peak ( E 15970). 4-Nitropyridine was ( 5 ) H. J. den Hertog, I;. Broekman and IT. Comb&, Rrr. trao. chiin.,

7 0 , 1 0 5 (1951).

\YILI,I.AM F. OLSZEWSKI

1701.

SI

prepared for spectral purposes from 4-nitropyridine-N-oxide according to the method of Ochiai.? The concentration of 4-nitropyridine-N-oxide throughout the course of the reaction was determined from the ultraviolet spectra by the application of Beer's law. A plot of the reciprocal of the concentration of 4-nitropyridine-N-oxide against time resulted in a straight line, indicating a second-order reaction ( k 0.0020 liter 'mole-sec.). In more detailed calculations it was shown that the effect of the small absorbancies of the other components a t 328 rnp was negligible.

Experimental Materials.-Quinoline-N-oxide dihydrate, 4-methoxypyridine-X-oxide and 4-nitropyridine-K-oxide (all Beacon Chemical Industries) were used as received. Pyridine-Koxide and 2-picoline-N-oxide (both Reilly Tar and Chemical Corp .) were distilled twice a t reduced pressure. 4-Picoline&--oxide (Reilly Tar and Chemical Corp.) was used as received. Triphenylphosphine (Eastman Kodak Co., White Label) and triethylene glycol (Eastman Kodak Co., Yellow Label) were used as received. Chromatographic sepamtions mere performed on .\lcoa alumina. Equipment.-Ultraviolet studies were carried out on a Beckman model DK-2 spectrophotometer using quartz cells with a 1-cm. path length. The absorption spectra were measured between 220 and 340 mp. General Procedure.-The reaction between triphenylphosphine and the amine oxides was performed by heating the reactants in a small distilling flask using a Wood metalbath to allow better temperature control. Pyridine, 2picoline and 4-picoline distilled from the reaction mixture a t the temperature indicated in Table I . The residue was recrystallized from a water-methanol solvent pair to obtain triphenylphosphine oxide. In the reactions of quinoline-N-oxide and 4-methoxypyridine-K-oxide with triphenylphosphine the products were isolated by column chromatography. After heating, the dark-colored reaction mixture was dissolved in a minimum amount of methanol, poured onto a chromatographic column, and eluted with petroleum ether (75-90') followed by cyclohexane. The free amine and triphenylphosphine oxide appeared in the early and late fractions, respectively. All reactions were carried out using approximately equimolar quantities of the amine oxide and triphenylphosphine. In each case the amine was identified by conversion t o its picrate. The concentration of solvent, lvhenever used, varied from 56-912'' by weight. Kinetic Study.-A mixture of 0.3548 g. (0.002532 mole) of 4-nitropyridine-?;-oxide and 0.6642 g. (0.002532 mole) of triphenylphosphine n a s placed in a 10-ml. volumetric flask which was nearlv filled with triethvlene glycol. The flask was inserted in 2 bath maintained at 1701f75' and as soon as the solids had melted the volume was brought up to the mark by adding more solvent. Then, 0.1-ml. aliquots were withdrawn and suitably diluted with methanol. Ultraviolet absorption spectra were determined for these solutions. The concentration of 4-nitropyridine-K-oxide was calculated by the application of Beer's law. Some transient absorption in the 4-nitro-pyridine region (282.5 mM) was ohserved, but this band disappeared during the course of tlic reaction. PIIIIADELPHIA 22, PESYA.